Nora Hunter,* James D. Foster, Grace Benson and James Hope. Introduction

Transcription

1 Journal of General Virology (1991), 72, Printed in Great Britain 1287 Restriction fragment length polymorphisms of the scrapie-associated fibril protein (PrP) gene and their association with susceptibility to natural scrapie in British sheep Nora Hunter,* James D. Foster, Grace Benson and James Hope Institute for Animal Health, AFRC/MRC Neuropathogenesis Unit, Ogston Building, West Mains Road, Edinburgh EH9 3JF, U.K. We have investigated the correlation between restriction fragment length polymorphisms of the scrapieassociated fibril protein (PrP) gene and the incidence of natural scrapie in British sheep during the period from July 1988 to November Sixty percent of the scrapie-positive animals studied were homozygous for a 6.8 kb EcoRI fragment (e 1) and a further 26 % carried el as heterozygotes. This fragment is linked to susceptibility to experimental scrapie in a closed flock of Cheviot sheep. Twelve percent of cases were found to be homozygous for a 4.4 kb EcoRI fragment (e3) which in the Cheviot flock has been linked to relative resistance to scrapie. A third EcoRI fragment of 5.2 kb (e2) has also been found but is relatively rare and has not yet been associated with scrapie susceptibility. Four sets of flocks affected by natural outbreaks of scrapie divided into two groups. In three of these flocks, all scrapie cases carried el with high frequencies of elel homozygotes. In the fourth, there were no elel scrapie cases; all scrapie sheep carried e3 in approximately equal numbers of heterozygotes and homozygotes. Introduction Scrapie, a disease of sheep and goats, is the archetype of a group of similar transmissible spongiform encephalopathies affecting mammals including Creutzfeldt-Jakob disease (CJD) of man and bovine spongiform encephalopathy (BSE) of cattle. Incubation periods of disease (or the appearance of clinical symptoms) have been shown to correlate with restriction fragment length polymorphisms (RFLPs) of the scrapie-associated fibril protein (PrP) gene in the case of experimental disease in mice (Westaway et al., 1987) and sheep (Hunter et al., 1989), and in the natural disease in man (Hsiao et al., 1989). This linkage and recent transgenic studies (Scott et al., 1989) suggest that the PrP gene itself may be the gene controlling the length of the incubation period. Apparently contradictory ideas on the genetic control of scrapie susceptibility have arisen from studies of flocks in which there was a high incidence of natural scrapie. Parry (whose lifetime's work was reviewed in Parry, 1984) did not believe that scrapie was naturally infectious and proposed that scrapie was exclusively a genetic disease, the result of an autosomal recessive gene. This work has its critics (Dickinson et al., 1965), and Gordon (1966), amongst others, proposed that the incidence of natural scrapie reflected inheritance of a recessive gene controlling susceptibility to scrapie although he questioned the full dominance of the 'resistance' allele. The influence of genotype on the susceptibility of sheep has been investigated by experimental infection. The Neuropathogenesis Unit (NPU) Cheviot sheep (a closed flock) are divided into two selected lines on the basis of their response to SSBP/1 scrapie (a standard source of scrapie infectivity and considered to be a mixture of several strains) (Dickinson, 1976). Sheep of one line (the positive line) develop scrapie following subcutaneous injection of SSBP/1 after about 300 days whereas sheep from the other (negative) line do not show symptoms within their normal lifetime. Negative line animals are not completely resistant to disease since intracerebral injection of SSBP/1 (and other sources of scrapie) will produce disease in these sheep after approximately 1000 days (Dickinson et al., 1968; Foster & Dickinson, 1988 a). The incubation period of scrapie in NPU Cheviot sheep is controlled by the gene Sip which has two alleles, sa and pa (Dickinson & Outram, 1988). Negative line sheep are Sip papa and positive line Sip sasa or sapa as the sa allele is thought to be dominant. Cross-breeding experiments (Foster & Dickinson, 1988b) using negative line NPU Cheviots and Suffolks from a flock bred for a high incidence of natural scrapie SGM

2 1288 N. Hunter and others (Dickinson et al., 1974; Foster & Dickinson, 1989) showed that the Sip gene also acted in controlling natural infection. However the fully recessive nature of the sa allele in response to natural infection has been questioned: there is a suggestion that heterozygotes may develop natural scrapie if infected early enough (Dickinson & Outram, 1988). Two RFLPs of the sheep PrP gene appear to act as markers for the Sip alleles in the NPU Cheviot flock (Hunter et al., 1989). With one of these (EcoRI), a fragment of 6.8 kb (el) is associated with Sip sa and a 4-4 kb fragment (e3) with Sip pa. The inability to identify the Sip genotype has hindered the study of natural disease susceptibility and pathogenesis. Therefore DNA from natural scrapie sheep was analysed for the polymorphic fragments detected with EeoRI to establish whether there was similar association with susceptibility. Methods Scrapie sheep. Blood and liver samples from scrapie sheep were obtained from Veterinary Investigation Centres (VICs) from various parts of Britain between July 1988 and November 1990 and a total of 113 scrapie sheep of 29 different breeds and cross-breeds were tested. We obtained 42 samples from Scotland, 62 from England and nine from Wales. In the majority of cases there was only one animal per flock but in one instance, five cases from a single flock were analysed (Group IV, see below). Clinical diagnoses of scrapie were confirmed by histopathology (carried out in the relevant VIC) and only confirmed cases were included in this study. Control sheep. The PrP gene RFLP frequency was studied in five groups of sheep to compare scrapie cases with unaffected animals and to obtain some idea of the RFLP type frequency in the national flock. Four of the sheep groups had already provided us with two or more scrapie cases; the fifth group was, until the time of writing, scrapiefree. Unaffected sheep were selected from the same flocks and of the same age range as the matched scrapie cases. During the course of these follow-up studies there were more scrapie cases in three groups. The extra cases were not added to the original study (of Table 2) since they were deliberately sought and might have biased the results. Group I is a Suffolk flock of approximately 200 sheep of which 37 healthy animals were studied and which, in 28 months, had four scrapie cases only three of which are in the original survey (of Table 2). Group II are Shetland sheep from flocks in Shetland which has tens of thousands of sheep. Fifty-eight unaffected animals from four different flocks were studied and the results were pooled as they were not markedly different from one another. These were then compared with 14 scrapie-affected flockmates only five of which were in the original survey. Group III are also Shetlands but from a flock in the north of England. It provided two scrapie cases in 28 months, both of which are in the original survey. These were compared with 12 unaffected flockmates. Group IV is a flock of about 400 sheep, most of which are Cheviot Border Leicester (Half bred) sheep. There have been 21 scrapie cases in this flock of about 400 animals during the study period and five of these were in the original survey. Thirty-five age-matched unaffected sheep from this flock were also RFLP-typed. Group V are Suffolk sheep from a flock at the East of Scotland College of Agriculture (ESCA). Sheep from 43 pedigree flocks throughout Britain were used to establish the flock between 1980 and In the hope of minimizing all disease and fertility risks, scrapie included, the chosen sheep were at least 6 years old. At the time of writing, no scrapie case has occurred in this flock. RFLP analysis. DNA was prepared from white blood cells or from liver tissue taken from scrapie or unaffected sheep using an Applied Biosystems Nucleic Acid Extractor (340A). For blood samples EDTA was used as an anti-coagulant because heparin inhibited restriction enzymes (data not shown). It was found that heparin could be removed by banding the DNA on isopycnic caesium chloride gradients and this had to be done for six natural scrapie samples and all the samples from control Group V. All other blood samples were collected into EDTA Vacutainers. Digestion, electrophoresis, blotting and hybridization of sheep DNA was carried out as described previously (Hunter et al., 1987). Nitrocellulose blots were probed with a sheep PrP genomic clone, pnpu42, a Sau3A subclone of pscr23.4 (Goldmann et at., 1990). Analysis of results. Results were analysed using chi-squared (~2) tests of association. Scrapie cases were compared with the whole group sample (i.e. scrapie sheep versus scrapie plus unaffected). Results NPU Cheviot flock The preliminary study of N PU Cheviot sheep (Hunter et al., 1989) was extended (Table 1 b) and revealed a small subgroup of the positive line (five out of 77 tested, 6.5 %) in which there was another polymorphism giving rise to a 5-2 kb EcoRI fragment (e2) (Table 1 a). None of our Cheviot sheep has so far been found to be homozygous for this fragment. As it occurs only in the positive line (Table 1 b and e), it may be a second relatively rare marker (along with el) for the Sip sa allele. However this is by no means certain and biological testing of sheep carrying this PrP gene type is being carried out to resolve the issue. Table l b shows the current PrP RFLP genotype information on 108 NPU Cheviot sheep (77 positive line, 31 negative line) taken over the period 1986 to At any one time there are approximately 100 sheep in each line. There is a clear association of the EcoRI RFLP type with scrapie susceptibility in this selected flock: PrP el is associated with relative susceptibility (Sip sa) to experimentally induced scrapie, and PrP e3 with relative resistance (Sip pa) (Table 1 c). However, as this would not necessarily be true in the national flock unselected for response to scrapie and subject to natural infection, a study of natural scrapie cases was undertaken. Natural scrapie Between July 1988 and November 1990, 113 scrapie samples were collected from VICs from all over Britain. The sheep were of 29 different breeds and crossbreeds; no individual breed had a markedly higher incidence of

3 Natural scrapie susceptibility 1289 Table 1. RFLPs of the sheep PrP gene in the NPU Cheviot flock* (a) Description of polymorphic types and association with Sip alleles EcoRI fragment size 6.8 kb 5-2 kb 4-4 kb RFLP type el e2 e3 Putative Sip allele sa 9 pa (b) EcoRI RFLP genotypes in NPU Cheviot flock Number of sheep Genotype elel ele2 ele3 e2e3 e3e3 Total Positive line Negative line (c) EcoRl RFLP type frequencies (~) in NPU Cheviot flock RFLP type el e2 e3 Putative Sip genotype Positive line sasa or sapa Negative line papa * (a) EcoRl restriction fragments found by Southern blot analysis of the NPU Cheviot flock. The linkage with the alleles of Sip (sa or pa) has been suggested by the analysis presented in (b) and (c). (b) Sheep from positive and negative lines with their corresponding PrP EcoRI RFLP genotypes in numbers of sheep. (c) Frequencies (%) of EcoRI RFLP types in positive and negative line sheep. Table 3. RFLP genotype analysis of the PrP gene in sheep affected b.v natural scrapie compared with unaffected flockmates* Number of each genotype Group of sheep n elel ele2 ele3 e2e2 e2e3 e3e3 X 2 P (a) Group I Scrapie Unaffected (b) Group II Scrapie Unaffected (c) Group III Scrapie 2 2 Unaffected (d) Group IV Scrapie Unaffected (e) Group V Unaffected < < > * EcoRI RFLP types are as designated in Table 1 (a). The sheep are as follows: (a) Group I, Suffolk sheep, Scotland; (b) Group II, Shetland sheep, Shetland; (c) Group III, Shetland sheep, north of England; (d) Group IV, Halfbred (Cheviot x Border Leicester) sheep, Scotland; (e) Group V, Suffolk sheep, ESCA, Edinburgh. n, Total number of sheep; X z and P calculated by comparison of scrapie sheep numbers with total sample numbers (total exposed sample). 19 Table 2. RFLPs of the PrP gene in natural scrapie-affected sheep* (a) Genotype elel ele2 ele3 e2e2 e2e3 e3e3 Total Number Frequency (~) (b) RFLP type el e2 e3 Number Frequency (~) * RFLP of the PrP gene in the sheep affected by natural scrapie. EcoRI RFLP types are as designated in Table 1 (a). scrapie. Age information was available for 102 sheep, the average being 3.2 years. This agrees well with previous surveys of the age-specific incidence of natural scrapie (Dickinson et al., 1964). The youngest affected sheep was a Charollais (1-2 years) and the oldest a Texel (7 years). There was no effect noted relating the age of sheep affected to the PrP polymorphic type (data not shown). Any apparent breed differences in age and/or PrP type were also thought not to be significant because of the small numbers involved. Full details of breeds and ages will be published elsewhere. PrP gene analysis of 113 scrapie sheep is presented in Table 2. Approximately 60~ (68) of 113 animals were PrP elel, 12~ (13) were PrP e3e3 and 25~ (28) were heterozygous. The other four animals carried PrP e2 (two were e2e2, with one each of ele2 and e2e3) (Table 2a). The RFLP type el had a frequency of approximately 73~ in these sheep; e3, 24~ and e2, 3~ (Table 2b). To assess the significance of these results, comparisons were made between the PrP genotypes of scrapie cases and age-matched unaffected flockmates. The five groups studied are described in detail in the Methods section. The scrapie cases listed in the Group studies are not all included in the results presented in Table 2. The Group study results (involving 250 sheep) are presented in Table 3 (genotypes) and Table 4 (RFLP type frequencies). Group studies (i) RFLP genotypes In Group I (a Suffolk flock) which compared four scrapie cases with 37 unaffected flockmates, there was an excess of PrP elel genotypes in the scrapie cases ( z= 11.52, P < 0.005) (Table 3a). PrP elel was a relatively rare genotype in the unaffected sheep sample. In Group II (Shetland sheep from four flocks) 14 scrapie sheep showed an excess of PrP elel (X 2 = 15.52, P < 0.005) compared with 58 healthy age-matched flockmates

4 1290 N. Hunter and others Table 4. RFLP type analysis of the PrP gene in sheep affected by natural scrapie compared with unaffected flockmates* Frequency (~) of each type Group of sheep n el e2 e3 X 2 P Group I Scrapie <0.005 Unaffected Group II Scrapie <0.005 Unaffected Group III Scrapie Unaffected Group IV Scrapie >0.995 Unaffected Group V Unaffected * EcoRl RFLP type frequencies in sheep groups from Table 3. n, Total number of sheep; X 2 and P calculated by comparison of scrapie sheep numbers with total sample numbers (total exposed sample). (Table 3b). There were no PrP e3e3 scrapie cases in Group I or Group II despite this genotype occurring at quite high frequencies in the healthy samples. In Group III (also a Shetland sheep flock but from the north of England) two scrapie cases were both PrP elel, but of the healthy flockmates 75 % (nine animals) were also PrP elel (Table 3c). This group was judged too small for meaningful analysis but it is interesting that, as with all the scrapie sheep in Groups I and II, the two scrapie sheep (the only cases in 28 months) in Group III carry el. Group IV appeared to be different. This Half bred sheep flock was chosen for further study because it had PrP e3e3 scrapie sheep cases, relatively uncommon (12%) in the original survey of Table 2. The 21 scrapie cases had very similar PrP RFLP genotypes to the 35 controls (X 2 = 2.46, P > 0-6) and all of the scrapie cases carried e3, approximately 50~ as homozygotes (Table 3d). Group V Suffolk sheep were selected from a flock partly selected to be scrapie-free: there has been no scrapie in this flock at the ESCA since it was set up in 1980 and until the time of writing. However when 67 sheep were tested with EeoRI, the genotype frequencies were found to be very similar to the unaffected Suffolks in the Group I flock (Table 3a) which is unselected and has had a scrapie problem. (ii) RFLP type frequencies Table 4 shows a comparison of the frequencies of e 1, e2 and e3 in scrapie-affected and unaffected sheep in all of the sheep groups studied. In Groups I and II PrP el is much more common in scrapie sheep than in healthy flockmates (X 2 = 8-66 and respectively, P < in both). The small Group III sample would suggest that el is present in the flock at high frequency (88%) and so it is not perhaps surprising that the two scrapie sheep should carry only el. The RFLP type frequencies in scrapie and unaffected sheep in Group IV are almost exactly the same (X z=0.007, P > 0.995). Group IV scrapie sheep have a higher frequency of e3 (71 ~) than do those from Groups I and II (12% and 11%), and a lower frequency of el (24%) compared with 88% (Group I) and 89% (Group II). In the unaffected sheep from these three groups, however, the RFLP type frequencies are similar with higher numbers of e3 than el. These results suggest that in Group IV scrapie is affecting sheep carrying e3 (common in the flock) and that in Groups I and II scrapie affects sheep carrying the less common el RFLP type preferentially. The unaffected sheep sample from the Suffolk (Group I) flock which is affected by scrapie was almost exactly the same in terms of RFLP type frequency (el, 32%; e3, 68%) as the scrapie-free Group V Suffock flock (el, 39%; e3, 61~). Discussion In this study of British sheep affected by natural scrapie, samples from 113 scrapie sheep of 29 breeds and crossbreeds taken from all over Britain during a 28 month period (up to November 1990) were tested for the frequency of an EcoRI PrP gene RFLP which has an association with scrapie susceptibility differences in NPU Cheviot sheep (Hunter et al., 1989). Eighty-six percent of the surveyed animals carried the PrP gene EcoRI fragment el which, in NPU Cheviot lines, has a clear association with the high susceptibility allele of Sip (sa). Sixty percent were homozygous (elel), 25% were ele3 and 1% ele2. The remaining animals were approximately 12% e3e3, 1% e2e3 and 2~ e2e2. In NPU Cheviots, e3 is associated with the low susceptibility allele of Sip (pa) but e2 has as yet uncertain association with alleles of Sip. Both elel (presumed Sip sasa) and ele3 (Sip sapa) Cheviots have relatively short incubation periods with SSBP/1 experimentally induced scrapie. The apparent association of e I with the incidence of natural scrapie was therefore encouraging. All the sheep included in this paper were also analysed with the enzyme HindlII, which also gives Sip-associated PrP gene RFLPs in the Cheviot sheep lines, but the use of this enzyme was not found to be generally informative. The group studies (which compare scrapie-affected sheep with age-matched unaffected flockmates) were carried out to provide some information on the RFLP type frequencies in the sheep population, to determine whether scrapie sheep had the same or different

5 Natural scrapie susceptibility 1291 genotypes as their flockmates and to try to establish whether the 12~ of scrapie animals with the RFLP genotype e3e3 were different from the others in the survey. When studying individual flocks which might have only two or three cases in 2 years, high numbers of scrapie sheep are difficult to obtain, except in the Group IV flock which has had 21 cases in 28 months and has more than equalled the numbers from Groups I, II and III combined. The numbers of sheep in the group study were therefore relatively small. The data analysis (chisquared tests of association) was carried out to give some indication of what is happening in these sheep groups but, because of the small numbers, should be treated with caution. The results from about 250 sheep have indicated that scrapie outbreaks may be divided into two sorts: one (the more common) in which scrapie affects predominantly PrP elel sheep, relatively less common in the flock (e.g. Groups I and II, Table 3a and b), and the other in which scrapie-affected sheep are not greatly different from their flockmates (e.g. Groups III and IV, Table 3c and d). In Group III the two scrapie sheep were also PrP elel but in Group IV none of the 21 cases was PrP elel : nine were PrP e3e3, 10 were PrP ele3 and two were e2e3. In Groups I, II and III, every scrapie sheep (20) carried el and in Group IV every scrapie sheep (21) carried e3 (Table 4). The RFLP type e2 was not found in Group I, II and III scrapie cases (although it was found at a frequency of 3 ~ in Group II unaffected sheep) but this RFLP type (e2) was found at a frequency of 5~ in scrapie-affected and unaffected sheep from Group IV. Comparison with the mouse experimental scrapie models is helpful in interpretation of these data. In mice the gene Sinc (homologue of Sip) controls the incubation period of all strains of scrapie. Like Sip, Sinc has two alleles, designated s7 and p7. With the ME7 strain of scrapie, Sinc s7s7 mice have a relatively short and Sinc p7p7 mice a prolonged incubation period. However with the 22A strain of scrapie this is reversed so that Sinc s7s7 mice have a longer incubation period than Sinc p7p7 (Dickinson & Outram, 1988). There are over 15 different strains of scrapie (Dickinson & Outram, 1988) and so any unknown mouse/scrapie strain combination with a short incubation period might involve any of the Sinc genotypes. The mouse PrP gene is linked to Sinc with at least six different RFLPs but the first of these to be tested, XbaI, was outside the coding region of the PrP gene and did not mark the Sinc alleles correctly in all the mouse strains tested (Hunter et al., 1987; Carlson et al., 1989). The sheep study of a non-coding region RFLP presented in this paper is at an equivalent stage to the early mouse work. It needed the discovery of a BstEII polymorphism within the mouse PrP-coding region to achieve 100~ linkage between Sinc and PrP (Westaway et al., 1987), and we are looking for a similarly strong marker within the coding region of the sheep PrP gene; one may have already been found in Suffolk sheep (Goldmann et al., 1990) but it has not yet been fully evaluated. There are at least two interpretations of our data on natural scrapie. The first assumes that there is only one kind of scrapie acting in the field and that only Sip sasa animals are susceptible, heterozygotes and Sip papa being much less susceptible or even resistant. In this case, the EcoRI el fragment may be helpful as a susceptibility marker: 60 ~ of cases examined so far were PrP elel and 26~ carried el as heterozygotes (Table 2). Therefore, 86~o of the scrapie sheep carried el. This may be similar to the early RFLP work in the mouse. Another interpretation is that, as in the experimental mouse-scrapie models, there is more than one strain of scrapie in the field and that the EcoRI RFLP genotype does indicate the correct Sip allele. Sheep developing scrapie from the more common strain could be Sip sasa (PrP elel) and less frequently Sip sapa (PrP ele3). This would mean that the sa allele is only partially dominant and we now have experimental evidence from SSBP/1 in NPU Cheviots to support this (Foster & Hunter, 1991). Following this line of argument, Sip papa (PrP e3e3) animals (for example Group IV Halfbred sheep, Table 3d) could be affected by a different strain of scrapie. [The Group IV flock had also by far the highest specific incidence of scrapie (cases/flock number) out of all the groups tested.] There is precedence for the existence of different types of sheep scrapie in the experimental scrapie source CH1641 which gives a higher incidence of disease in Sip papa NPU Cheviots than in the Sip sa carriers, the reverse of the SSBP/1 rankings (Foster & Dickinson, 1988a). SSBP/1 could therefore resemble the more common natural scrapie strain and CH1641 the more rare. This study involving more than 400 sheep has tested the idea that the apparent association between PrP RFLP genotypes and scrapie susceptibility is not limited to the NPU Cheviot flock but also applies to natural scrapie and to other breeds of sheep. The use of genetic linkage studies to provide markers for disease susceptibility has many pitfalls. However, the results so far look promising and natural scrapie may be more amenable to study than previously thought. The authors gratefully acknowlege the following: Veterinary Investigation Service staff from various centres in Scotland, Wales and England for providing samples, histopathology results and information; The Shetland Flock Health Association Ltd, William Dingwall (East of Scotland College of Agriculture); all farmers, shepherds and vets who supplied sheep material; and Professor W. G. Hill for assistance with data analysis.

6 1292 N. Hunter and others References CARLSON, G. A., GOODMAN, P. A., LOVETT, M., TAYLOR, B. A., MARSHALL, S. T., PETERSA-TORCHIA, M., WESTAWAY, D. & PRUSINER, S. B. (1989). Genetics and polymorphism of the mouse prion gene complex: control of scrapie incubation. Molecular and Cellular Biology 8, DICKINSON, A. G. (1976). Scrapie in sheep and goats. In Slow Virus Diseases of Animals and Man, pp Edited by R. H. Kimberlin. Amsterdam: North-Holland. DICKINSON, A. G. & OUTRAM, G. W. (1988). Genetic aspects of unconventional virus infections: the basis of the virino hypothesis. In Novel Infectious Agents and the Central Nervous System. Ciba Foundation Symposium No. 135, pp Edited by G. Bock & J. Marsh. Chichester: Wiley. DICKINSON, A. G., YOUNG, G. B., STAMP, J. T. & RENWICK, C. C. (1964). A note on the distribution of scrapie in sheep of different ages. Animal Production 6, DICKINSON, A. G., YOUNG, G. B., STAMP, J. T. & RENWlCK, C. C. (1965). An analysis of natural scrapie in Suffolk sheep. Heredity 20, DICKINSON, A. G., STAMP, J. T., RENWICK, C. C. & RENNIE, J. C. (1968). Some factors controlling the incidence of scrapie in Cheviot sheep injected with a Cheviot-passaged scrapie agent. Journal of Comparative Pathology 78, DICKINSON, A. G., STAMP, J. T. & RENWICK, C. C. (1974). Maternal and lateral transmission of scrapie in sheep. Journal of Comparative Pathology 84, FOSTER, J. D. & DICKINSON, A. G. (1988a). The unusual properties of CHl641, a sheep-passaged isolate of scrapie. Veterinary Record 123, 5-8. FOSTER, J. D. & DICKINSON, A. G. (1988b). Genetic control of scrapie in Cheviot and Suffolk sheep. Veterinary Record 123, 159. FOSTER, J. n. & DICKINSON, A. G. (1989). Age at death from natural scrapie in a flock of Suffolk sheep. Veterinary Record 125, FOSTER, J. D. & HUNTER, N. (1991). Partial dominance of the Sip gene in the control of experimental scrapie in Cheviot sheep. Veterinary Record (in press). GOLDMANN, W., HUNTER, N., FOSTER, J. n., SALBAUM, J. i., BEYREUTHER, K. & HOPE, J. (1990). Two alleles of a neural protein gene linked to scrapie in sheep. Proceedings of the National Academy of Sciences, U.S.A. 87, GORDON, W. S. (1966). Variation in susceptibility of sheep to scrapie and genetic implications. In Report of Scrapie Seminar, 1964 ARS U.S. Department of Agriculture 244. HSIAO, K., BAKER, H. F., CROW, T. J., POULTER, M., OWEN, F., TERWILLIGER, J. D., WESTAWAY, D., Oar, J. & PRUSINER, S. B. (1989). Linkage of a priori protein missense variant to Gerstmann- Straussler syndrome. Nature, London 338, HUNTER, N., HOPE, J., McCONNELL, I. & DICKINSON, A. G. (1987). Linkage of the scrapie-associated fibril protein (PrP) gene and Sinc using congenic mice and restriction fragment polymorphism analysis. Journal of General Virology 68, HUNTER, N., FOSTER, J. D., DICKINSON, A. G. & HOPE, J. (1989). Linkage of the gene for the scrapie-associated fibril protein (PrP) to the Sip gene in Cheviot sheep. Veterinary Record 124, PARRY, H. B. (1984). Scrapie. London: Academic Press. Scow, M., FOSTER, D., MIRENDA, C., SERBAN, D., CONFAL, F., WALCHLI, M., TORCHIS, M., GROTH, D., CARLSON, G., DEARMOND, S. J., WESTAWAY, D. & PRUSlNER, S. B. (1989). Transgenic mice expressing hamster priori protein produce species-specific scrapie infectivity and amyloid plaques. Cell 59, WESTAWAY, D., GOODMAN, P. A., MIRENDA, C. A., McKINLEY, M. P., CARLSON, G. A. & PRUSINER, S. B. (1987). Distinct prion proteins in short and long scrapie incubation period mice. Cell 51, (Received 5 December 1990; Accepted 19 February 1991)